I'm going WAY out on a limb here, because I'm not a genetics expert. But
here's my understanding of mtDNA. Where I'm wrong, I'd love to have someone
correct me.

Mitochondria are organelles within cells - that is, they appear almost to be
like little cells within cells - they have a cellular wall and their own
DNA - known as mtDNA. There are about 2000 Mitochondria in each cell, and
they process sugars into ATP - the "energy molecule" that is used in most
cellular processes.

When a cell divides, some of the mitochondria end up in each daughter cell.
The mitochondria multiply independently of cell division; thus a cell
divides, and each daughter cell has around 1000 mitochondria. As the
daughter cell's energy needs increase, the mitochondria divide within the
daughter cell, bringing the number of mitochondria up to "normal" levels.

Now let's move on to genetics. Sperm have few mitochondria, and during
fertilization, they are usually destroyed. Thus, the fertilized egg
contains nuclear DNA from both parents, but mitochondria (and hence mtDNA)
only from the mother.

In these mtDNA studies, they look at only portions of the mtDNA sequence.
In fact, they are especially interested in sections which DON'T code for
proteins, since mutations in that DNA are not effected by selection, and
should therefore occur at a moderately slow, STEADY rate. (A mutation to
the mtDNA that inhibited ATP production or other processes within the
mitochondrion would lead to the demise of the mitochondrion and the
mutation. A mutation to a non-coding region would neither result in the
demise of the organelle nor would it make the organelle more likely to
survive). Protein-coding sections of mtDNA can also be examined, though,
because many mutations have little or no effect on the synthesis of the
protein.

One key point is that mtDNA, being inherited only from the mother, stays
relatively constant generation to generation. A female and her offspring
will usually have THE SAME mtDNA. That's not true for nuclear DNA, where
the offspring's nuclear DNA is a blend of that of the parents. This makes
nuclear DNA much more variable within a population than mtDNA, and therefore
harder to study.

The study of mtDNA across populations make the primary assumption that
species *tend* to diverge in a tree-shaped pattern. That is, one species
splits into two or more species, each of which then splits into two or more
species, and so on. The aim of most of these studies is to determine the
relationship between the "leaves" on that tree - that is, the currently
existing species. If two species have recently diverged from the same
ancestral species, their mtDNA should be very similar. In fact, the mtDNA
of these "sister" species should be more similar than that of a "cousin"
species - one derived from, perhaps, a sister species of the ancestral
species. By assuming a steady mtDNA mutation rate, some measure of time can
even be inferred (i.e. "these species split X thousand years ago").

If speciation always occurred neatly, we'd have very little to talk about.
The classic model is that a species will get divided into two or more
populations that are [usually physically, but most importantly genetically]
isolated from each other. Over time, especially if the populations face
different conditions, these populations will diverge until they are distinct
species. Even if the populations then come back together, they will not
interbreed. Yeah, that's classic evolution.

Nature isn't so neat, though. Rather than a strictly branching tree of
species, there are often situations where the branches braid back together.
That is, before the two populations have diverged enough to become
completely separate, they can lose their genetic isolation from each other.
That's just one example of things that can happen that "violate" the
evolutionary tree-structure.

Another complicated scenario is a situation that leads to a cline. Imagine
that conditions vary north to south along the Pacific Coast in a fairly
linear fashion - conditions critical enough that the animals need to be
genetically adapted to latitude. Because the conditions vary linearly (that
is, they gradually change as you move north to south), any populations are
adapted such that they can thrive within a *fairly* broad area. For
instance, population A can thrive in the northernmost 500 miles, population
B can thrive in the next 500 miles south, and so on, through populations C,
D, and E. However, because conditions change linearly, an individual from
population A is not so different from an individual from population B; they
can interbreed, and their offspring may do well enough, especially around
the area exactly 500 miles south. An individual from population A that gets
blown way down south, however, is too different, and cannot thrive (and mate
successfully) among the residents of population E. Because *adjacent*
groups can freely interbreed, and because environmental conditions do not
usually act as selective pressure on mitochondria, the mtDNA can remain
fairly constant across all populations in the cline.

So with this new mtDNA "genetic bar coding" study, they found things that
don't fit with our current taxonomy, but they may be overstating their case.

First, pairs such as Golden-crowned and White-crowned Sparrows, and
Blue-winged and Cinnamon Teal, may simply have diverged too recently for
their mtDNA to have differentiated much. It doesn't mean they don't each
deserve species status. There is no minimum mtDNA difference threshold that
is required, since mtDNA difference simply (and crudely) denotes the amount
of *time* since divergence, and does not indicate the *degree* of nuclear
DNA divergence. (Not to mention that a measure of nuclear DNA divergence -
number of mutations or something - does not indicate the degree of
structural, behavioral, or social divergence).

Second, we have the postulated cline of Glaucous, Glaucous-winged, and
Western Gulls, moving north to south along the Pacific Coast. There seems
to be a lot of evidence that each is adapted to its latitudinal range, and
even that the hybrids between adjacent species are best adapted to the
narrow geographic region of hybridization. From this evidence, it is
probably correct to consider them distinct species, even if their mtDNA is
somewhat homogenous.

The cases of "hidden" species (I think Bewick's Wren was one) - where within
a single species there are populations whose mtDNA is different - well those
might well indicate that a split is in order. However, there are other
possible explanations. Perhaps these species do exist as isolated
populations. But perhaps, even though enough time has passed (while the
populations have been isolated) for their *mtDNA* to diverge, their nuclear
DNA may not have diverged.

If that last one seems strange, think of the horseshoe crab, where today's
crabs seem essentially identical to the horseshoe crabs from 300 million
years ago. They are apparently a stable organism over time, even though
their mtDNA is presumably mutating at the standard rate. Genetic changes
which result in structural change in the organism are apparently strongly
selected against, for some reason. Similarly, loons are often stated as
being one of the earliest families of birds, and have remained relatively
unchanged for millions of years. In contrast, the passerines have had an
explosion of "recent" speciation. These examples illustrate that some
organisms are more stable than others. Perhaps Bewick's Wrens are stable.
Perhaps they are dying out in the East because they cannot easily adapt to
changing conditions. If every mutation turns out to be undesirable, you're
stable but inflexible.

This is too long, but I'll send it anyway. If you've read this far, you
probably don't own a TV.

== Michael Hobbs
== Kirkland, WA
== http://www.marymoor.org/birding.htm
== birdmarymoor at verizon.net